Rectified Flow For Structure Based Drug Design
Daiheng Zhang, Chengyue Gong, Qiang Liu
TL;DR
The paper tackles structure-based drug design by reframing ligand generation as a rectified-flow transport conditioned on protein pockets. FlowSBDD leverages a velocity-field model to map an initial ligand distribution to a pocket-aware target, augmented with a bond-distance loss and flexible priors to improve learning and sample quality. On CrossDocked2020, FlowSBDD achieves state-of-the-art Avg. Vina Dock score $-8.50$ and $75.0\%$ Diversity, while offering faster sampling than diffusion-based methods. The method provides a flexible, scalable alternative to diffusion models, enabling targeted optimization through additional losses and priors with potential for practical impact in drug design.
Abstract
Deep generative models have achieved tremendous success in structure-based drug design in recent years, especially for generating 3D ligand molecules that bind to specific protein pocket. Notably, diffusion models have transformed ligand generation by providing exceptional quality and creativity. However, traditional diffusion models are restricted by their conventional learning objectives, which limit their broader applicability. In this work, we propose a new framework FlowSBDD, which is based on rectified flow model, allows us to flexibly incorporate additional loss to optimize specific target and introduce additional condition either as an extra input condition or replacing the initial Gaussian distribution. Extensive experiments on CrossDocked2020 show that our approach could achieve state-of-the-art performance on generating high-affinity molecules while maintaining proper molecular properties without specifically designing binding site, with up to -8.50 Avg. Vina Dock score and 75.0% Diversity.